Article pubs.acs.org/JPCA

Effect of Mutual Position of Electron Donor and Acceptor on Photoinduced Electron Transfer in Supramolecular Chlorophyll− Fullerene Dyads Kati Stranius,† Vladimir Iashin,‡ Taru Nikkonen,‡ Mikko Muuronen,‡ Juho Helaja,*,‡ and Nikolai Tkachenko*,† †

Department of Chemistry and Bioengineering, Tampere University of Technology, 33720 Tampere, Finland Department of Chemistry, University of Helsinki, 00100 Helsinki, Finland

‡

S Supporting Information *

ABSTRACT: In this study we have explored the influence of mutual position of chlorin electron donor and fullerene C60 electron acceptor on photoinduced electron transfer. Two zinc-chlorinaza-[18]crown-6 compounds and three pyrrolidino[60]fullerenes with alkyl aminium and varying coordinative moieties were synthesized and used for self-assembling of a set of complexes via twopoint binding. The aza[18]crown6 moieties were connected to chlorins via amide linker either at 134 or 174 position, hence, being attached on different sides of the chlorin plane. Furthermore, in the former case, the linker holds the crown closely spaced, whereas, in the latter, the linker gives more space and conformational freedom for the crown with respect to the chlorin macrocycle. The coordinative moieties at fullerene site, 3-pyridine, 4-pyridine, and 3-furan, were built by utilizing the Prato reaction. The two-point binding drove the molecules into specific complex formation by self-assembling; aminium ion was chelated by crown ether, while zinc moiety of chlorin was coordinated by pyridine and furan. Such pairing resulted in distinct supramolecular chlorin-fullerene dyads with defined distance and orientation. The performed computational studies at DFT level in solution, with TPSS-D3/def2-TZVP//def2-SVP, indicated different geometries and binding energies for the self-assembling complexes. Notably, the computations pointed out that for all the studied complexes, the donor−acceptor distances and binding energies were dictated by chirality of pyrrolidino ring at C60. The selective excitation of chlorin chromophore revealed efficient emission quenching in all dyads. The ultrafast spectroscopy studies suggested a fast and efficient photoinduced charge transfer in the dyads. The lifetimes of the charge separated states range from 55 to 187 ps in odichlorobenzene and from 14 to 60 ps in benzonitrile. Expectedly, the electron transfer rate was found to be critically dependent on the donor−acceptor distance; additionally, the mutual orientation of these entities was found to have significant contribution on the rate.

■

INTRODUCTION Chlorophylls play key roles in primary photosynthetic events, i.e., in light harvesting, energy transfer, and charge separation, occurring in natural photosystems.1,2 Elucidation of mechanistic details of these individual actions paves the way for development of artificial biomimetic devices such as solar light harvesting systems and photoactive molecular devices such as sensors, switches, and memories. Chlorophyll analogues, porphyrins, and phtalocyanines have been widely studied as electron donors.1,3−8 All these highly conjugated macrocycles absorb light over wide wavelength range in the visible and UV regions also possessing high electron−donor abilities. The central atom substitution and modification of the peripheral substituents of macrocycles can be used to change these properties for improving chromophore activity in donor−acceptor (DA) systems. Although nature has evolved chlorophyll to show excellence in the multitasking roles, the complicated purification of chlorophylls, instability of unmodified chlorophyll being exposure to light and limited amount of synthetic modification methods have limited its use in artificial systems.2,9−15 However, in comparison to the © 2014 American Chemical Society

semisynthetic chlorophyll analogues, the asymmetry of chlorophyll has its advantages over easily accessible symmetric synthetic porphyrines and phtalocyanines. It has been recently shown that the asymmetry enables precise supramolecular assemblies, e.g., derivation of chlorins for face-selective ligation16 and high order helix formation.17 One practical advantage of the supramolecular DA dyad selfassembling over much wider explored covalent dyad linking strategy is the possibility to prepare a series of donors and acceptors with the same interlocking groups and generate a large number of DA dyads solely by mixing different combinations of donor and acceptor molecules. Self-assembled systems are typically formed via metal−ligand or crown-ether coordination or ion pairing, hydrogen bonding or π−π stacking interactions.1,8,18−20 Self-assembling methods based on multiple modes of binding instead of one-point were also proposed and allowed to form complexes with higher binding strength and Received: December 19, 2013 Revised: February 3, 2014 Published: February 4, 2014 1420

dx.doi.org/10.1021/jp412442t | J. Phys. Chem. A 2014, 118, 1420−1429

The Journal of Physical Chemistry A

Article

better defined distances and orientation between the donor and acceptor.8 This is an important advantage of the multipoint binding, since the mutual orientation and distance of the functional units in DA dyads affect the electron transfer rate. In flexibly linked dyads, the distance and orientation vary and cannot be controlled.1,12,21−23 Fullerene C60 has high electron affinity, and electron transfer reactions involving fullerene are characterized by relatively small reorganization energy. Therefore, it is one of most commonly used acceptors in the DA design.24,25 The fullerene has low absorption in the visible region, which requires utilization of donor compounds with sufficient absorption in the visible range. Therefore chlorin−fullerene combination is an appropriate choice for supramolecular design. In this case the donor, chlorin, can be excited selectively, and the unique anion absorption band of the fullerene in the near-IR makes detection of the electron transfer dynamics reliable and reasonably accurate.22,26 In the present study, the supramolecular dyads were formed by specially designed Zn−chlorin and fullerene derivatives which implements two-point interaction to increase the binding efficiency and to achieve well determined mutual orientation between the donor and acceptor. Crown-ether moiety was attached to Zn−chlorin and alkyl aminium ion and pyridine entities to fullerene. The DA dyad was formed via axial coordination of pyridine entity to the zinc center of chlorin and complexation of the aminium ion with crown-ether. The photoinduced electron transfer in these dyads has been studied with steady-state and time-resolved spectroscopy.

Scheme 1. Chlorin-Derived Electron Donors (Chl-1 and -2) and C60-Derived Electron Acceptors (C60-1, -2, and -3)

rolog-2-111 fluorometer and corrected using the instrument response function supplied by the manufacturer. Solutions were prepared from solid compounds. Approximately 1 mg of each solid sample was dissolved in 1 mL of o-dichlorobenzene (ODCB) or benzonitrile (PhCN), and these solutions were used as stock solutions to prepare the samples for further measurements. Concentration of chlorophyll was maintained at ∼0.08−0.1 mM in all spectroscopic measurements. Due to low solubility of C60-2 at concentrations above 0.1 mM, special procedure was developed for preparing of C60-2 solution. One milligram of solid compound was first dissolved in 10 μL of dimethylformamide (DMF) and then 1 mL of ODCB was added to the solution. Time-Resolved Spectroscopy. Pump−probe and upconversion techniques were used to detect the fast photoinduced processes in solutions with a time resolution of 0.2 ps. The instruments and data analysis procedure have been described earlier.9,10,28 All measurements were carried out in 1 mm rotating cuvette. Absorbance of chlorophyll was around 0.8 (0.1 mM) at Soret band maximum and concentration of fullerene was adjusted to provide molecular ratio of roughly 1:1−5 (Chl@C60) depending on the emission quenching efficiency by the fullerenes. The excitation wavelength was around 420 nm, and the measurements were carried out in wavelength ranges 510−800 nm and 870−1100 nm. The excitation intensity was adjusted to provide 10−50% excitation efficiency. Differential Pulse Voltammetry. Differential pulse voltammetry (DPV) was used to measure the oxidation and reduction potentials of chlorins and fullerenes separately. Measurements were done at room temperature under nitrogen flow and recorded with a Iviumstat Compactstat IEC 61326 Standard potentiostat controlled by the Iviumsoft software. A platinum wire sealed in glass was used as working electrode and was polished with 0.05 μm alumina suspension before each

■

EXPERIMENTAL METHODS General. All the solvents used in synthesis and measurements were obtained as HPLC quality and used as received. 1H, 13 C, and COSY spectra were recorded at 27 °C using Varian Mercury 300 MHz or Varian Inova 500 MHz Spectrometer. 1H and 13C NMR spectra were referenced to the solvent signals (in CDCl3 7.26 and 77 ppm, respectively). High-resolution mass spectra were obtained on HR-ESI MS with Bruker micrOTOF instrument. Compounds. The structures of the electron donating chlorins (Chl-1 and Chl-2) and the electron accepting pyrrolidino-[60]fullerenes (C60-1, C60-2, C60-3) are presented in Scheme 1. Chl-1 was synthesized by EDCI-assisted coupling of pyropheophorbide a with 1-aza-18-crown-6 followed by insertion of zinc (see Supporting Information (SI)). Chl-2 was prepared via selective aminolysis of 134position, which has been recently developed in our laboratory,27 and subsequent zinc insertion (see SI). Fullerenes C60-1, C60-2, and C60-3 were synthesized via Prato reaction27 and further one-pot deprotection of amine moiety and aminium salt formation (see SI). It should be noted that each synthesized pyrrolidino[60]fullerene was obtained as a 1:1 mixture of two enantiomers (R and S). The asymmetry is invisible in a symmetrical environment for instance when similar mixtures of enantiomers have been complexed with symmetric porphyrin derivatives.22,23 However, in case of chiral chlorophyll derivatized host molecules each enantiomer results in different DA distances and geometries. This will be discussed in detail later. Steady State Spectroscopy. The UV−vis spectral measurements were carried out with a Shimadzu UV-3600 UV−vis−NIR spectrophotometer. The fluorescence spectra were recorded with an ISA-Jobin Yvon-SPEX-Horiba Fluo1421

dx.doi.org/10.1021/jp412442t | J. Phys. Chem. A 2014, 118, 1420−1429

The Journal of Physical Chemistry A

Article

The aim of this work was time-resolved study of the photoinduced electron transfer in supramolecular dyads, which was carried out in 1 mm thick cuvettes and sample absorption close to 1 at the Soret band. Therefore, titration experiments were done in conditions close to those used in pump−probe measurements, i.e., in 1 mm cuvettes and chlorin concentrations roughly 0.1 mM. Figure 2 shows changes in absorption

measurement. Ag/AgCl wire was used as reference electrode and platinum wire as a counter electrode. Measurements were done in 0.1 M TBAClO4 in ODCB (dried and stored under molecular sieves) as supporting electrolyte and ferrocene was used as an internal reference redox system. Concentration of all measured samples was 0.46 mM. Scanning rate was 0.05 V/s, the pulse time 20 ms, the pulse amplitude 20 mV, and the equilibration time 5s. The potential cycling was performed between −1 and 1 V and vice versa. Computational Methods. Computational study was performed at DFT level in solution for all the self-assembling pairs including the both R- and S-enantiomers of fullerene. All computations were performed using the Turbomole 6.4 program package.29 Solvation effects were taken into account using COSMO solvation model in all computations with dielectric constant of o-dichlorobenzene (ε = 10).30 A keyword “use_contcav” was used to assure that no COSMO cavity was built inside fullerene ring. The TPSS-D331,32 functional was chosen with double ζ quality basis set, def2-SVP, for optimizations, and single point properties were calculated using triple ζ quality basis set, def2-TZVP.33 The empirical dispersion correction D3 was used with zero-damping. The MARI-J approximation was used in all computations with suitable auxiliary basis set.34−37 To discover the effect of the functional and basis set, Chl-1@S-C60-3 complex was also optimized using TPSS-D3/def2-TZVP and B97-D3.32,38 Change of the functional did not lead to any notable changes in the structure and upgrade of the basis set did not change nature of the result. Therefore, all optimizations were performed using def2-SVP basis set. Grid m4 was used throughout all computations. Pictures were generated using TmoleX, and orbitals were plotted using isovalue of 0.01.

Figure 2. Corrected titration absorption spectra of increasing concentration of C60-1 in the solution of Chl-2 (0.1 mM) in ODCB. The red-brown line represents the starting point with no C601 present.

spectrum of Chl-2 upon addition of C60-1 in ODCB. Titration was done by adding the fullerene solution to chlorin solution so that the sample volume was increased by 15% at the end of titration. The absorption spectra presented in Figure 2 were corrected by taking the changes in chlorin concentration into account. The concentration of fullerene increases during the titration to the level at which fullerene absorption cannot be ignored, and therefore no clear isosbestic points can be seen in the figure. Observed diminished intensities and red shifts of the Soret and first Q bands along with increasing intensities of other Qy-bands indicate the binding of C60-1. Absorption spectra of the Chl-2@C60-1, -2, and -3 supramolecular dyads (Figure 1 and Figure S16 in SI) show 1−3 nm red shifts and decrease in intensities of the bands compared to the free Chl-2, which is typical for ligand coordination to zinc atom in the chlorin macrocycle.3 The chlorin band shift upon formation of Chl-1@C60-1, -2, and -3 dyads is less pronounced being less than 1 nm, though the decrease in intensity is also well observed. To get better insight to the complex formation, we studied the fluorescence quenching of the both chlorins, Chl-1 and Chl-2, in the presence of C60-1, -2, and -3. The best wavelength for selective excitation of chlorin chromophore in the dyad was found to be 615 nm which corresponds to the second intense Qy-band. Fluorescence spectra of Chl-2 mixtures with C60-1, -2, and -3 in ODCB are shown in Figure 3. The fluorescence intensity decreases upon addition of fullerene, but the shape of the emission spectrum remains intact. This is typical behavior for emission quenching via energy or electron transfer mechanism in dyads. The most efficient quenching (80%) was observed for Chl-2@C60-1 complex already at 1:1 ratio. For other complexes (Table 1, Figure S17 in the SI) the emission quenching was less than 50% for 1:1 ratios and therefore higher amounts of fullerenes were needed for achieving higher quenching at concentrations suitable for pump−probe measurements. With an excess of C60 weakest quenching (less than 60%) was recorded for both

■

RESULTS Steady State Spectroscopy. The normalized absorption spectra of the both chlorins, Chl-1 and Chl-2, and 3-pyridine derivatized pyrrolidino[60]fullerene, C60-1, in ODCB are presented in Figure 1. The absorption spectra of Chl-2 show intense Soret band at 430 nm and Qy band at 662 nm. There is a minor (less than 2 nm) difference in positions of the bands for Chl-1 compared to Chl-2. The ratio of Soret and Q-band intensities is also somewhat different in the studied chlorins as can be expected considering different substituents at position 134.

Figure 1. Absorption spectra of Chl-1, Chl-2, and C60-1 in ODCB. Concentrations of Chl-2 and C60-1 are 0.1 mM, while absorption spectrum of Chl-1 is normalized to match that of Chl-2 at the Soret band. 1422

dx.doi.org/10.1021/jp412442t | J. Phys. Chem. A 2014, 118, 1420−1429

The Journal of Physical Chemistry A

Article

calculated complex concentration is lower than that determined from the fluorescence quenching. This can be due to possible formation of complexes when fullerene is bound to two chlorin units by pyridine and crown sites respectively. However, Job’s method measurements carried out at concentration 0.1 mM indicated formation of 1:1 complexes (Figure S18 in the SI), which confirms that in conditions used in time-resolved measurements, the dominating complex stoichiometry is 1:1. The values of K for each dyad are listed in Table 1. The magnitude of the K value for Chl-1 and Chl-2 are in agreement with that previously measured for a similar dyad,12 suggesting enhanced binding efficiency due to two-point binding. The combination Chl-2@C60-1 has even 1 order of magnitude larger equilibrium constant than others. In general, C60-1 and C60-2 form strong supramolecular dyads with both chlorins. Binding constants for furan functionalized C60-3 were found to be lower compared to pyridine derivatives C60-1 and C60-2. Emission quenching was also studied in PhCN. Figure 4 shows the emission spectra for Chl-2 and its 1:3 mixture with

Figure 3. Relative emission spectra Chl-2@C60 mixtures in ODCB with ratios used for time-resolved measurements. Concentration of Chl-2 is 0.1 mM in all samples.

Table 1. Chl@C60 Concentration Ratios (Chlorin Concentration 0.1 mM), Measured Emission Quenching Values (I0 − I)/I0 and Binding Constants (K) for Samples Used for Time-Resolved Measurements in ODCB chlorin

fullerene

ratio

(I0 − I)/I0

Chl-1 Chl-1 Chl-1 Chl-2 Chl-2 Chl-2

C60-1 C60-2 C60-3 C60-1 C60-2 C60-3

1:3 1:3 1:5 1:1 1:3 1:5

0.70 0.80 0.55 0.80 0.70 0.55

K, M−1 1.6 7.8 4.4 4.6 6.8 4.7

× × × × × ×

104 104 103 105 104 103

chlorin−C60-3 complexes (1:5). While for Chl-2@C60-2 and both Chl-1@C60-1 and Chl-1@C60-2 over 70% quenching was achieved at 1:3 ratio. The limiting factor in these cases was relatively poor solubility of C60-2 and C60-3, which reduces accuracy and reliability of the measurements. Also absorption of the fullerene chromophores at the excitation wavelength, 615 nm, could not be neglected anymore, and it was accounted for in evaluation of the quenching efficiency in Table 1 and Figure S16 in the SI. To improve the solubility of C60-2, a small amount of DMF (10 μL) was added to ODCB for dissolving fullerene at higher concentrations. There was no notable difference in solubility of C60-3 with such amount of DMF, which did not affect the properties of the dyads otherwise. Accounting for the fullerene absorption at the excitation wavelength gives estimation for actual quenching lower than 60% even at 1:5 ratios for C60-3 complexes, and thus one can expect that not more than half of chlorin molecules form the dyads in the best case. A usual method to determine binding constants, K, is to present the emission quenching data in Stern−Volmer plot, and calculate the constant from the slope of data linear approximation. However, the linear dependence is only observed when the concentration of the emitting compound, chlorin, is much smaller compared to that of the quencher, fullerene, which is not the case of this study, since the binding constants have to be determined at concentrations close to those used in transient absorption measurements. Therefore, an exact model for the complex formation (see SI) was used to fit the dependence of the fluorescence intensity on the concentration of the compounds to obtain the binding constants (Figure S17 in the SI). It was noted that for Chl2@C60-1 at low concentration of C60-1 (when the concentration of chlorin is higher than that of fullerene) the

Figure 4. Emission spectra of Chl-2 and its 1:3 mixture with C60-1 in PhCN. Excitation wavelength was 615 nm. Inset in plot shows absorption spectra of measured samples. Concentration of Chl-2 is 0.1 mM in all samples.

C60-1. The emission quenching at ratio 1:1 was 26%, and the 1:3 ratio was needed for both chlorins to achieve 50% quenching (Figure S19 in the SI). Thus the emission quenching in PhCN was found to be less efficient compared to that in ODCB. Electrochemical Studies. Differential pulse voltammetry (DPV) measurements were carried out to estimate the oxidation and reduction potentials of the compounds (Figure 5). The first oxidation potentials of Chl-1 and Chl-2 were located at 0.08 and 0.10 V vs Fc/Fc+, respectively. The oxidation was reversible only for the latter compound. The second oxidation potentials were located at 0.33 and 0.45 V vs Fc/Fc+, for Chl-1 and Chl-2, respectively, but were not reversible for either of the chlorins. The first reduction of the fullerene derivatives was located at −1.17 V vs Fc/Fc+. The difference between chlorin first oxidation and fullerene first reduction potentials gives a rough estimation of the energy of the charge separate state. The energies are 1.25 and 1.27 eV for Chl-1 and Chl-2 based dyads, respectively, without accounting for the Coulombic interaction. Measured oxidation potentials for Chl-2 correspond well to that of a similar Zn-Chlorin.39 The oxidation potential of Chl-1 is slightly shifted by 20 mV, which is likely a consequence of the crown ether orientation over the aromatic chlorin ring. 1423

dx.doi.org/10.1021/jp412442t | J. Phys. Chem. A 2014, 118, 1420−1429

The Journal of Physical Chemistry A

Article

Figure 5. Differential pulse voltammograms of (a) Chl-1 and Chl-2 (0.45 mM) and (b) C60-1 (0.45 mM) in 0.1 M TBAClO4/ODCB.

Time-Resolved Measurements. The fluorescence decay measurements were done for all the self-assembling pairs with the fluorescence up-conversion technique. The decays were recorded at 670 nm, which is roughly the maximum of the steady-state fluorescence of chlorin chromophore. The samples, 1:1−5 mixtures of chlorin and fullerene derivatives, were excited at 420 nm. At this wavelength, chlorin absorption is higher than that of fullerene even for 1:5 mixtures (Figure S20 in the SI). Furthermore, since the emission quantum yield of fullerene is much lower than that of chlorin, the contribution of the fullerene fluorescence at the monitoring wavelength can be neglected and the measured emission decays can be attributed to chlorin solely. Figure 6 shows the fluorescence decays for

Table 2. Time Constants Obtained from Up-Conversion and Pump−Probe Measurements for the Chl@C60 Dyads in ODCB and PhCN chlorin

fullerene

measured in ODCB: Chl-1 C60-1 Chl-1 C60-2 Chl-1 C60-3 Chl-2 C60-1 Chl-2 C60-2 Chl-2 C60-3 measured in PhCN: Chl-1 C60-1 Chl-2 C60-1 Chl-2 C60-2 a

τa, ps

τCS, ps

τCR, ps

2.0 2.5 3.1 1.6 2.2 4.4

1.7 0.3

63 76

5 7

55 187

n.a. 2.6 7

14 15 60

Fluorescence lifetime obtained from up-conversion measurements.

spectrum range from 500 to 1100 nm, and fitted globally. At least a three-exponential decay model has to be used to achieve a reasonable goodness of the global fit of the pump−probe data. Transient absorption decay component spectra and timeresolved spectrum right after excitation (t = 0 ps) obtained from three-exponential global fit of pump−probe data of Chl2@C60-1 (1:1) are presented in Figure 7a. The long-lived component (lifetime >2 ns) arises from nonbonded chlorins (Figure S21 in the SI), and its relative intensity depends on the efficiency of the DA complex formation. The shorter-lived components were observed for all DA combinations. They originate from the interaction between the chlorin and fullerene counterparts, and, they are the subject of further analysis. The 55 ps component in Figure 7a has spectrum typical for the CS state with bleaching of the ground-state absorption of chlorin, new band around 750 nm, and also broad absorption in the NIR region with a band at 1040 nm. It is known that the fullerene anion has a transient absorption band around 1000 nm and chlorin cation near 750 nm.15,40−42 Therefore this state was attributed to the charge separated (CS) state with lifetime of 55 ps. The fastest component (5 ps) shows formation of the CS state Chl+−C60− as indicated by the band formation at 1040 nm (negative intensity of the component) which is typical for the fullerene anion. At the same time in the Q-band region the component shows an increase in the bleaching of the ground state of chlorin. This behavior resembles that of covalently linked phytochlorin-fullerene dyads,9 in which the formation of charge separated state was mediated by an exciplex state and appeared in two competing pathways. The first includes

Figure 6. The fluorescence decays of Chl-1 with C60-1, C60-2, and C60-3 monitored at 670 nm in ODCB. Concentration of Chl-1 is 0.1 mM in all samples, and Chl-1@C60 concentration ratios are indicated in the plot.

Chl-1 dyads in ODCB. The emission decays consist of two clear components. The fast components in the picoseconds time domain were attributed to the emission decays in supramolecular dyads, and the calculated lifetimes are summarized in Table 2. The longer components have lifetimes in nanosecond time domain, and attributed to free chlorin molecules. The lifetime of the singlet excited state in supramolecular dyads increases in order C60-1 < C60-2 < C60-3 for both chlorin derivatives. The proportion of the fast to slow components is in agreement with emission quenching efficiency. The highest quenching was observed with C60-1, which also shows highest proportion of the fast to slow components. The transient absorption measurements (pump−probe) were carried out with excitation at the same wavelength as for emission decays, 420 nm. The data were collected in a wide 1424

dx.doi.org/10.1021/jp412442t | J. Phys. Chem. A 2014, 118, 1420−1429

The Journal of Physical Chemistry A

Article

Figure 7. Transient absorption decay component spectra of the (a) Chl-2@C60-1 (1:1) and (b) Chl-1@C60−1 (1:3) in ODCB. Concentration of chlorins is 0.1 mM.

molecular complexes (Table 1). Similar spectral features were also seen for other dyads (Figure 7b and S22 in SI). The lifetimes for CS states are listed in Table 2. However, due to low binding constant of C60-3 with both Chl-1 and Chl-2, it was not possible to resolve the CS state lifetime from pump− probe measurements, and for that reason only lifetimes of the first singlet state (from up-conversion measurements) are provided in Table 2. In PhCN solution, a weaker complex formation was observed for all the self-assembling pairs. Accordingly, the lower relative dyad population was concomitant with detection of less intense CS signal. Nonetheless, the spectral main features of the studied samples were fundamentally similar in both ODCB and PhCN. The transient absorption decay component spectra obtained from four-exponential global fit of pump−probe data of Chl-2@C60-1 (1:3) in PhCN are presented in Figure 8. The

formation of the exciplex from the singlet excited state of chlorin chromophore and the exciplex relaxation to the CS state. The second starts from the energy transfer from the primary excited chlorin to the fullerene and follows by the exciplex formation and its relaxation to the complete charge separated state. The latter route results in the recovery of the ground state bleaching of the chlorin upon the energy transfer and following bleaching as the exciplex and CS states are formed. The spectrum of the 5 ps component can be interpreted in a similar manner. For part of the complexes, the primary reaction is the energy transfer from chlorin to fullerene, and when the CS takes place starting from the fullerene singlet excited state the reinvolvement of the chlorin in the relaxation pathway is observed as an increase of bleaching at the Q-band. The degree of recovery of the Q-band bleaching can serve as a rough measure of the relative part of dyads, which pass through the fullerene singlet excited state, and for this dyad it was found to be ≤25%. There is a noticeable difference in the time constants of the fast component obtained for Chl-2@C60-1 from the transient absorption and emission decay measurements, 5 and 1.7 ps, respectively (Table 2). Apparently, the excited singlet state does not decay directly to the CS state, and many other processes contribute to that what is observed as a 5 ps component in transient absorption measurements. These are energy transfer, electron transfer starting from both the chlorin and fullerene excited state, and possibly exciplex intermediate prior to formation of the complete CS state. In fluorescence decay measurements, only the chlorin singlet excited state is observed, which is the initial state for all of the following reactions. Attempts to resolve more intermediates in transient absorption measurements were not successful: four-exponential fit gave only 5% improvement in the sigma value and a relatively short-lived component (∼0.3 ps) with the spectrum indicating some changes in the Q-band region and around 1000 nm but not allowing unambiguous assignment of the component. The accuracy of the transient absorption measurements was insufficient to identify all processes preceding the formation of the CS state. One reason for this is the limited yield of complexes under experimental conditions favorable for the transient absorption measurements, which was found to be 80% in the best case. Therefore we will limit our following discussion by the analysis of the lifetime of the CS state of different combinations of the donors and acceptors, which was determined with reasonable accuracy. For samples other than Chl2@C60-1 higher amount of fullerene was needed to achieve sufficient yield of supra-

Figure 8. Transient absorption decay component spectra of the Chl2@C60-1 (1:3) in PhCN. Concentration of Chl-2 is 0.1 mM.

components with time constants 2.6 and 15 ps can be identified as those corresponding to charge separation and recombination, respectively. Compared to the same dyad in ODCB, the lifetime of CS state is 4 times shorter in PhCN, as can be expected since higher solvent polarity lowers the energy of the charge separate state and higher reorganization energy, which cumulatively results in shorter lifetime of the CS state in the case of chlorin−fullerene dyads.9 Similar spectral features were also observed for Chl-2@C60-2. However, in the case of Chl1@C60-1, the signal was weak and only component corresponding to the charge recombination was obtained (Figure S23 in the SI). The lifetimes for CS states in PhCN are presented in Table 2. The order of CS state lifetimes is the 1425

dx.doi.org/10.1021/jp412442t | J. Phys. Chem. A 2014, 118, 1420−1429

The Journal of Physical Chemistry A

■

same both in ODCB and PhCN. Lifetimes of complexes with C60-2 were found to be higher compared to those with C60-1. DFT Calculations. In order to acquire some theoretical insight about complex formations DFT calculations were performed with DFT TPSS-D3/def2-TZVP//def2-SVP method for all the complexes (Figure 9). Photophysically critical

distance parameters, i.e., chlorin−fullerene center-to-center and edge-to-edge spacing, and energy differences between the fullerene derivative enantiomers are listed in Table 3. Table 3. Computed Geometry Parameters for the Chl@C60 Dyads fullerene

RE−E, Å

RC−C, Å

Erel, kcal/mol

Chl-1

R-C60-1 S-C60-1 R-C60-2 S-C60-2 R-C60-3 S-C60-3 R-C60-1 S-C60-1 R-C60-2 S-C60-2 R-C60-3 S-C60-3

3.2 3.4 6.5 3.4 3.3 3.3 4.2 3.5 6.3 4.0 4.2 3.7

8.9 8.9 8.9 10.9 8.9 9.2 9.6 8.6 9.6 10.7 9.6 8.8

0.00a 2.56 0.00 −1.97a 0.00a 5.15 0.00 −3.49a 0.00a 7.07 0.00 −3.94a

Chl-1 Chl-1 Chl-2 Chl-2 Chl-2 a

DISCUSSION

Some important remarks can be drawn from the absorption spectra upon titration. The most pronounced differences in the absorption spectra can be seen in Chl-2 upon formation of complexes with C60-1 and C60-2. Titration of Chl-2 with furan derivative C60-3 does not show remarkable changes in absorption, due to the fact that the oxygen electron pair donating ability is lower compared to that of nitrogen, which makes it a looser ligand for Zn compared to pyridine coordinating group. In the case of Chl-1, relatively small changes are observed while titrating it with C60-1, -2, or -3. Structurally, the crown moiety of Chl-1 is hanging in the terminal position of long flexible propionic amide side arm (174), meanwhile in Chl-2, it is attached into a close proximity of chlorin macrocycle at the 132-position. The latter seems to fit better to the geometry of the fullerene counterparts. Considering two enantiomers of fullerene, a simple equilibrium model may be insufficient to explain the observed results, and the latter may differ at different concentrations. At low concentration of chlorin and sufficiently high concentration of fullerene, one can expect that most of complexes will be formed by enantiomer providing lower energy of the complex, and the logarithm of the complex concentration ratio to be proportional to the energy difference divided by the thermal energy, kT. This is the limit of low chlorin concentration and high excess of fullerenes, and obviously it is not suitable for the time-resolved spectroscopy studies. At higher chlorin concentration and chlorin:fullerene ratio close to 1:1, the outcome depends on the equilibrium constants for each pair that each chlorin forms either with S- or R-C60 enantiomer. If products of the binding constants and chlorin (or fullerene) concentrations for both enantiomers are greater than one, both enantiomers will form the complexes, leaving almost no free chlorins or fullerenes. In this case, we expect to see high fluorescence quenching efficiency as is the case of Chl-2@C601 in conditions used in this study (80%, Table 1). If the product is greater than one only for a lower energy complex, only the lower energy complex will be formed predominantly, and fluorescence quenching will be roughly 50%, assuming that enantiomers present in solution in equal amounts. This is the case of Chl-1@C60-1 and Chl-2@C60-2, and, to a lesser degree, of Chl-1@C60-2. To increase the yield of complexes, we increased the concentration of fullerene, and under this condition, again, the lower energy complex will be formed predominantly since it has higher binding constant, and the enantiomers are competing in the complex formation again. The latter is the case of the pump−probe measurements for all samples except Chl-2@C60-1. It should be noted that the equilibrium constants were determined using standard methods, which do not account for presence of enantiomers, and thus the determined constants are not accurate and correspond to the lower energy complexes. On the basis of the above discussion, we can presume that in three cases of time-resolved spectroscopy measurements, we studied monoenantiomer complexes, namely, Chl-1@R-C60-1, Chl-1@S-C60-2 and Chl-2@R-C60-2. In the case of Chl-2@ C60-1, we probably have minor domination of Chl-2@S-C601, but the properties of Chl-2@S-C60-1 and Chl-2@R-C60-1 seem to be rather close to each other, and we cannot distinguish between them within experimental accuracy. The combination Chl-2@C60-1 has 1 order of magnitude larger equilibrium constant compared to those of other studied

Figure 9. Geometry-optimized structures with depicted empty and filled frontier orbitals for (a) Chl-2@R-C60−2 and (b) Chl-2@SC60−1 (LUMO = blue-gray and HOMO = dark-light green) with isosurface value 0.01. In calculations, CF3COO− counterion is replaced by Cl− ion (yellow sphere).

chlorin

Article

Complex formation is more favorable.

Expectedly, diastereomeric chlorin derivatives showed enantiospesific binding and energetics for the fullerene enantiomers. The calculated energies of complexes for both R- and Sfullerene enantiomers show that the relative energy of Chl-2@ S-C60-1 is lower than that of Chl-2@R-C60-1 (Table 3). Thus, it is apparent that in excess of racemic C60-1 formation of Chl2@S-C60-1 complex dominates over that of Chl-2@R-C60-1. Additionally, the calculations also insist that in case of Chl-2@ C60-2 complex R-enantiomer of C60-2 is more favorable toward binding, whereas for Chl-2@C60-3 S-enantiomer is expected to be bound stronger. However, in case of Chl-1 the trend is opposite: R-C60-1, S-C60-2, and R-C60-3 are clearly favored in the analogous complex formations. Overall, the computational studies suggest that both Chl-1 and Chl-2 have an ability to discriminate between different enantiomers of studied fullerene derivatives, although racemic mixtures of C601, C60-2, and C60-3 were used for all the measurements. 1426

dx.doi.org/10.1021/jp412442t | J. Phys. Chem. A 2014, 118, 1420−1429

The Journal of Physical Chemistry A

Article

favorable for each chlorin diastereomer with particular fullerene enantiomer. Thus, in solution of a racemic mixture of fullerene derivatives, enantioselective complexes Chl-2@S-C60-1, Chl2@R-C60-2, Chl-2@S-C60-3, Chl-1@R-C60-1, Chl-1@SC60-2 and Chl-1@S-C60-3 were most likely formed, which is also supported by the experimentally observed behavior of studied systems. The time-resolved absorption spectroscopy studies demonstrated fast charge separation with time constants less than 10 ps in supramolecular dyads. The lifetimes of the charge separated states are in 55−190 ps range in nonpolar ODCB and 15−60 ps in polar PhCN depending on the DA distance in the dyads and giving the damping factor β = 0.44 Å−1 for the distance dependence. Compared to previously measured covalently bonded chlorin-fullerene dyads, the results showed that mutual orientation of donor and acceptor in dyads also influences the electron transfer rates. Namely, at the same edgeto-edge distance, the electronic coupling is stronger for fullerene located on top of macrocycle than at the macrocycle periphery.

complexes. Thus it is the best matching pair for ditopic binding. Another advantage of ditopic binding is the complex formation in both relatively nonpolar ODCB and polar PhCN. Since the proportion of the charge recombination time constants is the same in these two solvents, we can presume that the complex geometries are rather similar in two solvents.22 The differences in the lifetimes of CS state for different combination of donors and acceptors in supramolecular dyad can be attributed to different geometries of the dyads and can be discussed further based on the results of the dyad modeling. In a simplified case the lifetime distance dependence is given by an exponential law, τCR ∼ exp(−βr), where r is the distance and β is a damping factor, and it arises from the exponential decay of the electronic coupling with the distance.43 An open question here is how to measure the distance r. Two usual approaches are to look at the center-to-center distance, RC−C, and edge-to-edge distance, RE‑E. From Table 2 one can pick up two pairs of dyads that have photophysically similar features, namely Chl-2@S-C60-1 and Chl-2@R-C60-2, and Chl-1@RC60-1 and Chl-1@S-C60-2, to examine whether these complexes have also some structural relations. For the first pair, the lifetimes differ 3.4 times, and for the latter by 1.2 times only. For the first pair, the difference in RC−C is 1 Å, and for the second it is 2 Å, which does not agree with the lifetime differences. However, for the former pair, the difference in RE−E is 2.8 Å, and for the latter pair it is 0.2 Å, which supports very well the observed difference in the charge recombination time constants and suggests the damping factor to be 0.44 Å−1. The damping factors reported for through-space electron transfer for a number of porphyrin-based DA systems are in the range 0.3−0.6 Å−1.44−46 Therefore, one may conclude that the edgeto-edge distance is the key parameter for the distance dependence. On the other hand, the CS state lifetime for covalently linked Zn phytochlorin-fullerene dyad was reported to be 21 ps in benzonitrile,9 which is longer than that of Chl1@R-C60-1 and Chl-2@S-C60-1 dyads (14 and 15 ps, respectively), despite the edge-to-edge distance being shorter. In this case, center-to-center distance correlates better with the lifetimes. Striking difference in geometries between the mentioned covalent-linked dyad and the supramolecular dyads in this study is that in covalently linked dyad the fullerene is located closer to the peripheral part of the chlorin macrocycle, whereas for the supramolecular dyads, the fullerenes are on top of the macrocycle. One explanation is that in the latter case, the electronic coupling is stronger at the same edge-to-edge distance, and the center-to-center distance appears to be better value to compare the lifetimes. In both cases dyads with C60-1 (3-pyridine linker) had shorter CS state lifetime than that with C60-2 (4-pyridine linker). This is the expected result based on the calculated distances between donor and acceptor, which showed that the DA distance is shorter for dyads with C60-1 acceptor than that with C60-2 acceptor.

■

ASSOCIATED CONTENT

S Supporting Information *

Synthesis of molecules, UV/vis and emission titration spectra, equations and fittings for binding constant determination, timeresolved transient absorption spectra and detailed information about computational methods. This material is available free of charge via the Internet at http://pubs.acs.org

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: juho.helaja@helsinki.fi (J.H.). *E-mail: nikolai.tkachenko@tut.fi (N.T.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work has been supported by Academy of Finland [No. 135113 and 135058]. Dr. Petri Heinonen is accredited for the HRMS measurements. The National Centre for Scientific Computing (CSC) is acknowledged for computational resources. V.I. and T.N. acknowledge the Graduate School of Organic Chemistry and Chemical Biology (GSOCCB). V.I. is also grateful to Magnus Ehnrooth and University of Helsinki foundations for financial support.

■

REFERENCES

(1) Wróbel, D.; Graja, A. Photoinduced Electron Transfer Processes in Fullerene−Organic Chromophore Systems. Coord. Chem. Rev. 2011, 255, 2555−2577. (2) Grimm, B.; Porra, R. J.; Rüdiger, W.; Scheer, H. Chlorophylls and Bacteriochlorophylls: Biochemistry, Biophysics, Functions and Applications; Springer: Dordrecht, The Netherlands, 2006. (3) D’Souza, F.; Ito, O. Photoinduced Electron Transfer in Supramolecular Systems of Fullerenes Functionalized with Ligands Capable of Binding to Zinc Porphyrins and Zinc Phthalocyanines. Coord. Chem. Rev. 2005, 249, 1410−1422. (4) Araki, Y.; Ito, O. Factors Controlling Lifetimes of Photoinduced Charge-Separated States of Fullerene-Donor Molecular Systems. J. Photochem. Photobiol., C 2008, 9, 93−110. (5) El-Khouly, M. E.; Ito, O.; Smith, P. M.; D’Souza, F. Intermolecular and Supramolecular Photoinduced Electron Transfer Processes of Fullerene−Porphyrin/Phthalocyanine Systems. J. Photochem. Photobiol., C 2004, 5, 79−104.

■

CONCLUSIONS Two zinc chlorins equipped with aza-[18]crown-6 ether moiety at 134 or 174 positions (Chl-2 and Chl-1, respectively) and three pyrrolidine functionalized fullerenes armed with alkyl aminium and 3-furan, 3- or 4-pyridine linkers (C60-3, C60-1, and C60-2, respectively) were synthesized to study selfassembled DA dyads and characterized by computational and spectroscopic methods. The computational studies revealed that the complex formation was energetically clearly more 1427

dx.doi.org/10.1021/jp412442t | J. Phys. Chem. A 2014, 118, 1420−1429

The Journal of Physical Chemistry A

Article

Type Interactions on Photochemical Charge Stabilization in “TwoPoint” Bound Supramolecular Porphyrin−Fullerene Conjugates. Chem.Eur. J. 2005, 11, 4416−4428. (23) D’Souza, F.; Chitta, R.; Gadde, S.; McCarty, A.; Karr, P.; Zandler, M.; Sandanayaka, A.; Araki, Y.; Ito, O. Design, Syntheses, and Studies of Supramolecular Porphyrin−Fullerene Conjugates, Using Bis-18-Crown-6 Appended Porphyrins and Pyridine or Alkyl Ammonium Functionalized Fullerenes. J. Phys. Chem., B 2006, 110, 5905−5913. (24) Ito, O.; D’Souza, F. Recent Advances in Photoinduced Electron Transfer Processes of Fullerene-Based Molecular Assemblies and Nanocomposites. Molecules 2012, 17, 5816−5835. (25) Imahori, H.; Sakata, Y. Fullerenes as Novel Acceptors in Photosynthetic Electron Transfer. Eur. J. Org. Chem. 1999, 1999, 2445−2457. (26) Guldi, D. M.; Prato, M. Excited-State Properties of C60 Fullerene Derivatives. Acc. Chem. Res. 2000, 33, 695−703. (27) Maggini, M.; Scorrano, G.; Prato, M. Addition of Azomethine Ylides to C60: Synthesis, Characterization, and Functionalization of Fullerene Pyrrolidines. J. Am. Chem. Soc. 1993, 115, 9798−9799. (28) Isosomppi, M.; Tkachenko, N.; Efimov, A.; Lemmetyinen, H. Photoinduced Electron Transfer in Double-Bridged Porphyrin− Fullerene Triads. J. Phys. Chem., A. 2005, 109, 4881−4890. (29) TURBOMOLE V6. 4; University of Karlsruhe and Forschungszentrum Karlsruhe GmbH, 1989−2007; TURBOMOLE GmbH, s. 2., 2012. (30) Schafer, A.; Klamt, A.; Sattel, D.; Lohrenz, J. C. W.; Eckert, F. COSMO Implementation in TURBOMOLE: Extension of an Efficient Quantum Chemical Code Towards Liquid Systems. Phys. Chem. Chem. Phys. 2000, 2, 2187−2193. (31) Tao, J.; Perdew, J. P.; Staroverov, V. N.; Scuseria, G. E. Climbing the Density Functional Ladder: Nonempirical Meta-Generalized Gradient Approximation Designed for Molecules and Solids. Phys. Rev. Lett. 2003, 91, 146401. (32) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent and Accurate Ab Initio Parametrization of Density Functional Dispersion Correction (DFT-D) for the 94 Elements H-Pu. J. Chem. Phys. 2010, 132, 154104. (33) Weigend, F.; Ahlrichs, R. Balanced Basis Sets of Split Valence, Triple Zeta Valence and Quadruple Zeta Valence Quality for H to Rn: Design and Assessment of Accuracy. Phys. Chem. Chem. Phys. 2005, 7, 3297−3305. (34) Sierka, M.; Hogekamp, A.; Ahlrichs, R. Fast Evaluation of the Coulomb Potential for Electron Densities Using Multipole Accelerated Resolution of Identity Approximation. J. Chem. Phys. 2003, 118, 9136−9148. (35) Eichkorn, K.; Weigend, F.; Treutler, O.; Ahlrichs, R. Auxiliary Basis Sets for Main Row Atoms and Transition Metals and Their Use to Approximate Coulomb Potentials. Theor. Chem. Acc. 1997, 97, 119−124. (36) Weigend, F. Accurate Coulomb-Fitting Basis Sets for H to Rn. Phys. Chem. Chem. Phys. 2006, 8, 1057−1065. (37) Hellweg, A.; Hättig, C.; Höfener, S.; Klopper, W. Optimized Accurate Auxiliary Basis Sets for RI-MP2 and RI-CC2 Calculations for the Atoms Rb to Rn. Theor. Chem. Acc. 2007, 117, 587−597. (38) Grimme, S. Semiempirical GGA-Type Density Functional Constructed with a Long-Range Dispersion Correction. J. Comput. Chem. 2006, 27, 1787−1799. (39) Nonomura, Y.; Igarashi, S.; Yoshioka, N.; Inoue, H. Spectroscopic Properties of Chlorophylls and their Derivatives. Influence of Molecular Structure on the Electronic State. Chem. Phys. 1997, 220, 155−166. (40) Tkachenko, N. V.; Rantala, L.; Tauber, A. Y.; Helaja, J.; Hynninen, P. H.; Lemmetyinen, H. Photoinduced Electron Transfer in Phytochlorin-[60] Fullerene Dyads. J. Am. Chem. Soc. 1999, 121, 9378−9387. (41) El-Khouly, M. E.; Araki, Y.; Fujitsuka, M.; Watanabe, A.; Ito, O. Photoinduced Electron Transfer between Chlorophylls (a/b) and

(6) Guldi, D. M.; Zilbermann, I.; Gouloumis, A.; Vázquez, P.; Torres, T. Metallophthalocyanines: Versatile Electron-Donating Building Blocks for Fullerene Dyads. J. Phys. Chem., B. 2004, 108, 18485− 18494. (7) Schuster, D. I.; Li, K.; Guldi, D. M.; Palkar, A.; Echegoyen, L.; Stanisky, C.; Cross, R. J.; Niemi, M.; Tkachenko, N. V.; Lemmetyinen, H. Azobenzene-Linked Porphyrin−Fullerene Dyads. J. Am. Chem. Soc. 2007, 129, 15973−15982. (8) Bottari, G.; Trukhina, O.; Ince, M.; Torres, T. Towards Artificial Photosynthesis: Supramolecular, Donor−Acceptor, Porphyrin- and Phthalocyanine/Carbon Nanostructure Ensembles. Coord. Chem. Rev. 2012, 256, 2453−2477. (9) Tkachenko, N. V.; Rantala, L.; Tauber, A. Y.; Helaja, J.; Hynninen, P. H.; Lemmetyinen, H. Photoinduced Electron Transfer in Phytochlorin-[60]Fullerene Dyads. J. Am. Chem. Soc. 1999, 121, 9378−9387. (10) Vehmanen, V.; Tkachenko, N. V.; Imahori, H.; Fukuzumi, S.; Lemmetyinen, H. Charge-Transfer Emission of Compact Porphyrin− Fullerene Dyad Analyzed by Marcus Theory of Electron-Transfer. Spectrochim. Acta, Part A 2001, 57, 2229−2244. (11) Vehmanen, V.; Tkachenko, N. V.; Efimov, A.; Damlin, P.; Ivaska, A.; Lemmetyinen, H. The Role of the Exciplex State in Photoinduced Electron Transfer of Phytochlorin-[60]Fullerene Dyads. J. Phys. Chem., A 2002, 106, 8029−8038. (12) Kavakka, J. S.; Heikkinen, S.; Kilpelainen, I.; Tkachenko, N. V.; Helaja, J. Zn Pyro-Pheophorbide a-Fulleronicotine Dyad; Supramolecular Self Assembled Donor-Acceptor System for Photoinduced Charge Separation. Chem. Commun. 2009, 7, 758−760. (13) Zheng, G.; J. Dougherty, T.; K. Pandey, R.; K. Pandey, R. Novel Chlorin-Diene Building Block by Enyne Metathesis: Synthesis of Chlorin−Fullerene Dyads. Chem. Commun. 1999, 24, 2469−2470. (14) Montforts, F.; Kutzki, O. Simple Synthesis of a Chlorin− Fullerene Dyad with a Novel Ring-Closure Reaction. Angew. Chem., Int. Ed. 2000, 39, 599−601. (15) Fukuzumi, S.; Ohkubo, K.; Imahori, H.; Shao, J.; Ou, Z.; Zheng, G.; Chen, Y.; Pandey, R. K.; Fujitsuka, M.; Ito, O.; Kadish, K. M. Photochemical and Electrochemical Properties of Zinc Chlorin-C60 Dyad as Compared to Corresponding Free-Base Chlorin-C60, FreeBase Porphyrin-C60, and Zinc Porphyrin-C60 Dyads. J. Am. Chem. Soc. 2001, 123, 10676−10683. (16) Kavakka, J. S.; Heikkinen, S.; Helaja, J. Zn Pyropheophorbide a: A β-Face Selective Nicotine Receptor. Eur. J. Org. Chem. 2008, 2008, 4932−4937. (17) Shinozaki, Y.; Richards, G.; Ogawa, K.; Yamano, A.; Ohara, K.; Yamaguchi, K.; Kawano, S.; Tanaka, K.; Araki, Y.; Wada, T.; Otsuki, J. Double Helices of a Pyridine-Appended Zinc Chlorophyll Derivative. J. Am. Chem. Soc. 2013, 135, 5262−5265. (18) D’Souza, F.; Maligaspe, E.; Sandanayaka, A. S. D.; Subbaiyan, N. K.; Karr, P. A.; Hasobe, T.; Ito, O. Photochemical Charge Separation in Supramolecular Phthalocyanine−Multifullerene Conjugates Assembled by Crown Ether-Alkyl Ammonium Cation Interactions. J. Phys. Chem. A 2010, 114, 10951−10959. (19) El-Khouly, M. E.; Kay, K.; D’Souza, F.; Fukuzumi, S. Supramolecular Tetrad of Subphthalocyanine−Triphenylamine−Zinc Porphyrin Coordinated to Fullerene as an “Antenna-Reaction-Center” Mimic: Formation of a Long-Lived Charge-Separated State in Nonpolar Solvent. Chem.Eur. J. 2010, 16, 6193−6202. (20) Rodríguez-Morgade, M. S.; Plonska-Brzezinska, M.; Athans, A. J.; Carbonell, E.; de Miguel, G.; Guldi, D. M.; Echegoyen, L.; Torres, T. Synthesis, Characterization, and Photoinduced Electron Transfer Processes of Orthogonal Ruthenium Phthalocyanine−Fullerene Assemblies. J. Am. Chem. Soc. 2009, 131, 10484−10496. (21) D’Souza, F.; Deviprasad, G.; Zandler, M.; El-Khouly, M.; Fujitsuka, M.; Ito, O. Photoinduced Electron Transfer in “Two-Point” Bound Supramolecular Triads Composed of N,N-Dimethylaminophenyl−Fullerene−Pyridine Coordinated to Zinc Porphyrin. J. Phys. Chem., A 2003, 107, 4801−4807. (22) D’Souza, F.; Chitta, R.; Gadde, S.; Zandler, M.; McCarty, A.; Sandanayaka, A.; Araki, Y.; Ito, O. Effect of Axial Ligation Or π−π1428

dx.doi.org/10.1021/jp412442t | J. Phys. Chem. A 2014, 118, 1420−1429

The Journal of Physical Chemistry A

Article

Fullerenes (C60/C70) Studied by Laser Flash Photolysis. Photochem. Photobiol. 2001, 74, 22−30. (42) Kesti, T. J.; Tkachenko, N. V.; Vehmanen, V.; Yamada, H.; Imahori, H.; Fukuzumi, S.; Lemmetyinen, H. Exciplex Intermediates in Photoinduced Electron Transfer of Porphyrin−Fullerene Dyads. J. Am. Chem. Soc. 2002, 124, 8067−8077. (43) Wasielewski, M. R. Photoinduced Electron Transfer in Supramolecular Systems for Artificial Photosynthesis. Chem. Rev. 1992, 92, 435−461. (44) Imahori, H.; Tamaki, K.; Guldi, D. M.; Luo, C.; Fujitsuka, M.; Ito, O.; Sakata, Y.; Fukuzumi, S. Modulating Charge Separation and Charge Recombination Dynamics in Porphyrin−Fullerene Linked Dyads and Triads: Marcus-Normal Versus Inverted Region. J. Am. Chem. Soc. 2001, 123, 2607−2617. (45) Kang, Y. K.; Rubtsov, I. V.; Iovine, P. M.; Chen, J.; Therien, M. J. Distance Dependence of Electron Transfer in Rigid, Cofacially Compressed, π-Stacked Porphyrin-Bridge-Quinone Systems. J. Am. Chem. Soc. 2002, 124, 8275−8279. (46) Wiberg, J.; Guo, L.; Pettersson, K.; Nilsson, D.; Ljungdahl, T.; Mårtensson, J.; Albinsson, B. Charge Recombination Versus Charge Separation in Donor−Bridge−Acceptor Systems. J. Am. Chem. Soc. 2007, 129, 155−163.

1429

dx.doi.org/10.1021/jp412442t | J. Phys. Chem. A 2014, 118, 1420−1429